Mouse models of human cancer: Noninvasive phenotyping with MRI

Thursday, October 11, 2012 — Poster Session III
10:00 a.m. – Noon	Natcher Conference Center, Building 45	NCI	IMAG-13

Authors

• SA Jansen
• Y Song
• L Ileva
• L Lu
• M Anver
• P Martin
• T Van Dyke

Abstract

What can noninvasive imaging reveal about the underlying biology of cancer? We have established a clinically relevant framework for image-based characterization of two distinct cancers in mice, brain astrocytoma and breast ductal carcinoma. We applied this approach in over 200 mice representing a diversity of mouse models, from commonly used xenografts to advanced GEMMs wherein key molecular pathways relevant to human astrocytoma (Rb,Ras,PTEN) and breast cancer (Rb,p53) are genetically altered. Furthermore, we developed a new high-resolution radiologic-pathologic correlation technique to accurately co-register even the earliest stages of cancer (~100microns). To facilitate clinical applicability, a key element of our approach is to acquire and analyze murine magnetic resonance imaging (MRI) data analogously to human data. This includes T1 FFE, T2 TSE and DCEMRI obtained at 3Tesla. We took the novel step of adapting two human MRI lexicons for application in mice: the BIRADS (breast) and VASARI (brain) descriptors. In doing so we found that, unlike xenografts, GEMMs recapitulated the heterogeneity of the human MRI phenotype. This study establishes a new strategy for image-based characterization of cancer in mouse models, including acquisition, analysis and pathologic correlation. With this framework, we can embark on further investigation into the biological underpinnings of image-based features.

back to top